The Yield Curve Inversion
The spread between the 10-year and 2-year US Treasury yields has inverted before seven of the last eight US recessions since 1968, giving it an 87.5% accuracy rate. The lag between inversion and recession onset has historically ranged from 7 to 24 months. The lone exception was 1966, when the curve inverted without a recession following.
The problem is what happened recently. The 2022-2023 inversion lasted 16 months (July 2022 to November 2023), making it the longest inversion in modern history. Despite this extended duration, no recession materialized as of late 2025. This is a significant challenge to the indicator's reliability.
Using the yield curve alone is like having an alarm with a variable delay between the signal and the event. It might go off 7 months before the problem or 24 months before, or, as we have now seen, it might not correspond to a recession at all. Context matters. Recent research suggests the inversion alone is not sufficient; house prices must decline and corporate credit spreads must significantly increase for a recession to follow.
The Sahm Rule
Claudia Sahm, a former Federal Reserve economist, developed a simple real-time recession indicator. When the three-month moving average of the unemployment rate rises by 0.5 percentage points or more from its low over the prior 12 months, the economy is in or entering a recession.
In July 2024, the Sahm Rule was triggered when the three-month average reached 0.53 percentage points above its 12-month low. But Sahm herself publicly stated she did not believe the economy was in recession. In a Bloomberg opinion piece on August 7, 2024, she explained that post-pandemic labor supply distortions, particularly immigration-driven increases in the labor force, pushed unemployment higher without the demand weakness that normally accompanies a Sahm Rule trigger. The unemployment rate rose because more people entered the labor force, not because fewer people had jobs.
This was the indicator's first false positive, and its creator was the one who called it out. It is a useful reminder that even indicators with perfect historical records operate on assumptions that can break down when underlying economic structures change.
The Conference Board Leading Economic Index
The LEI combines 10 economic indicators including initial jobless claims, building permits, manufacturer new orders, and the S&P 500. It is designed to turn down before recessions. Historically, the LEI has declined for 3-6 consecutive months before every post-war recession.
The LEI declined for an extended period in 2022-2023 without an official recession. Part of the issue is that the index gives heavy weight to manufacturing indicators, which can decline while the larger services sector remains strong. The composition of the economy has shifted substantially since the index weights were originally calibrated.
The Composite Approach
No single indicator is reliable enough to use in isolation. Each captures a different facet of the economy and has different lead times, false positive rates, and blind spots. The practical approach is to track multiple indicators simultaneously and look for convergence.
When the yield curve inverts, the LEI declines for consecutive months, initial jobless claims start rising, credit spreads widen, AND the Sahm Rule triggers, you have a convergence that is historically very reliable. When only one or two indicators flash while the others remain healthy, the signal is ambiguous.
A composite scorecard tracking five to seven recession indicators, each with its own trigger threshold and historical accuracy, gives you a much stronger read on economic conditions than following any single metric. It also helps you distinguish between genuine deterioration and the periodic false alarms that make any individual indicator less useful than its track record suggests.